CN113895271A - Deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method - Google Patents
Deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method Download PDFInfo
- Publication number
- CN113895271A CN113895271A CN202111203154.7A CN202111203154A CN113895271A CN 113895271 A CN113895271 A CN 113895271A CN 202111203154 A CN202111203154 A CN 202111203154A CN 113895271 A CN113895271 A CN 113895271A
- Authority
- CN
- China
- Prior art keywords
- current charging
- direct
- early warning
- data
- power direct
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000012544 monitoring process Methods 0.000 title claims abstract description 36
- 238000013135 deep learning Methods 0.000 title claims abstract description 22
- 241000287127 Passeridae Species 0.000 claims abstract description 42
- 238000013136 deep learning model Methods 0.000 claims abstract description 20
- 238000004458 analytical method Methods 0.000 claims abstract description 13
- 238000010845 search algorithm Methods 0.000 claims abstract description 13
- 238000011156 evaluation Methods 0.000 claims abstract description 8
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 25
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 206010000369 Accident Diseases 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 230000006399 behavior Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000002789 length control Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000013450 outlier detection Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 claims description 2
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 claims description 2
- 230000002457 bidirectional effect Effects 0.000 claims description 2
- 238000012549 training Methods 0.000 claims description 2
- 238000013461 design Methods 0.000 abstract description 2
- 230000008901 benefit Effects 0.000 description 5
- 239000000178 monomer Substances 0.000 description 3
- 238000012300 Sequence Analysis Methods 0.000 description 2
- 229910052799 carbon Inorganic materials 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L53/00—Methods of charging batteries, specially adapted for electric vehicles; Charging stations or on-board charging equipment therefor; Exchange of energy storage elements in electric vehicles
- B60L53/60—Monitoring or controlling charging stations
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L3/00—Electric devices on electrically-propelled vehicles for safety purposes; Monitoring operating variables, e.g. speed, deceleration or energy consumption
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L3/00—Electric devices on electrically-propelled vehicles for safety purposes; Monitoring operating variables, e.g. speed, deceleration or energy consumption
- B60L3/12—Recording operating variables ; Monitoring of operating variables
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/70—Energy storage systems for electromobility, e.g. batteries
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/7072—Electromobility specific charging systems or methods for batteries, ultracapacitors, supercapacitors or double-layer capacitors
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02T90/10—Technologies relating to charging of electric vehicles
- Y02T90/12—Electric charging stations
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02T90/10—Technologies relating to charging of electric vehicles
- Y02T90/16—Information or communication technologies improving the operation of electric vehicles
- Y02T90/167—Systems integrating technologies related to power network operation and communication or information technologies for supporting the interoperability of electric or hybrid vehicles, i.e. smartgrids as interface for battery charging of electric vehicles [EV] or hybrid vehicles [HEV]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S30/00—Systems supporting specific end-user applications in the sector of transportation
- Y04S30/10—Systems supporting the interoperability of electric or hybrid vehicles
- Y04S30/12—Remote or cooperative charging
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Development (AREA)
- Sustainable Energy (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The invention designs a deep learning-based method for monitoring the high-power direct-current charging state and early warning faults of an electric vehicle, which comprises the following steps of firstly, monitoring the state of various parameters of the high-power direct-current charging of the electric vehicle, and storing the parameters into a database; secondly, dividing data in the database into historical data and real-time data, and preprocessing the historical data and the real-time data; then, designing a CNN-BiGRU deep learning model to fully learn the normal direct current charging historical data, constructing a direct current charging prediction model of the electric automobile, and optimizing the hyper-parameters of the model by adopting a sparrow search algorithm; then, an evaluation standard of the model prediction precision is formulated to evaluate the accuracy of the model prediction, residual error analysis is carried out on the model prediction value through a sliding window method, and a fault early warning threshold value and a fault early warning rule which are suitable for direct current charging of the electric automobile are determined; and finally, applying the trained CNN-BiGRU prediction model to real-time high-power direct-current charging monitoring of the electric automobile to realize fault early warning of the electric automobile.
Description
Technical Field
The invention belongs to the technical field of equipment fault early warning, and particularly relates to a deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method.
Background
The electric automobile can relieve the energy crisis, reduce carbon emission and protect the environment, meets the target of the double-carbon strategy, and is a key development object of governments and enterprises of all countries at present. With the rapid development of electric vehicles, the endurance mileage of electric vehicles is increasing, so that the power battery is difficult to be fully charged in a short time. The high-power direct-current charging of the electric automobile has the advantage of greatly shortening the charging time of the electric automobile, and is widely applied to the field of electric automobile charging, so that the safety and reliability of the high-power direct-current charging of the electric automobile are emphasized in the industry. The power source of the electric automobile is mainly various batteries, once the batteries have safety problems in the charging process, the electric automobile is likely to have a fire, and irreparable economic loss and even personnel injury are caused. Therefore, the online monitoring of the state of the power battery of the electric automobile under high-power direct current charging is very important for fault early warning before the electric automobile generates a fire accident.
At present, research results on electric vehicles at home and abroad are many, but the content on high-power direct-current charging early warning of the electric vehicles is relatively less. The Convolutional Neural Network (CNN) can extract the advantage of deep features in the high-power direct-current charging data of the electric vehicle, can make full use of the information of the charging data of the electric vehicle, and improves the accuracy of the charging prediction model. The bidirectional gating circulation Unit BiGRU (Bi-directional Gated Current Unit, BiGRU) can simultaneously consider the characteristics of historical and future electric vehicle high-power direct-current charging data, can fully and deeply utilize the electric vehicle high-power direct-current charging data information, and enables the model to have stronger data extraction, analysis and generalization capabilities.
In order to ensure the charging safety of the electric automobile, effectively predict the occurrence of charging accidents in time and prevent spontaneous combustion accidents of the electric automobile in the high-power direct-current charging process, the patent obtains the high-power direct-current charging data of the electric automobile through national standard standards and provides a method for monitoring the high-power direct-current charging state of the electric automobile and early warning faults of the electric automobile based on deep learning by utilizing the obtained data. Firstly, CNN is used for fully utilizing the charging historical data of the monitored electric automobile, deep features hidden in the charging data are extracted, the advantages of BiGRU analysis history and future data are utilized, time sequence analysis is carried out on the extracted deep features, and a high-power direct-current charging prediction model of a normal electric automobile is built. Secondly, searching for hyper-parameters such as the learning rate, the iteration times and the number of hidden units of the CNN-BiGRU deep learning model by using a sparrow search algorithm so as to enhance the prediction accuracy of the direct-current charging prediction model. Then, an evaluation standard of the model prediction accuracy is established, and the prediction result of the charging prediction model is evaluated. And then, analyzing and processing the prediction residual error of the charging model by adopting a sliding window analysis method, and determining a good fault early warning threshold value. And finally, applying the charging prediction model meeting the requirements and the determined early warning threshold value to the real-time monitoring of the high-power direct-current charging of the electric automobile, and realizing the fault early warning of the high-power direct-current charging process of the electric automobile.
Disclosure of Invention
The invention provides a method for monitoring the high-power direct-current charging state of an electric vehicle and early warning faults based on deep learning, aiming at the safety problem in the high-power direct-current charging process of the electric vehicle, and aiming at solving the defects of the existing method. According to the method, firstly, CNN is used for carrying out deep mining on normal charging historical data of the electric automobile, deep features of the CNN are extracted, BiGRU is used for carrying out full analysis and utilization on the deep features, and a charging prediction model of the electric automobile is constructed. Secondly, searching for hyper-parameters such as iteration times, the number of hidden units and the like of the CNN-BiGRU deep learning model by using a sparrow search algorithm so as to enhance the prediction accuracy of the direct-current charging prediction model. Then, an evaluation standard of the model prediction accuracy is established, and the prediction result of the charging prediction model is evaluated. And then, analyzing and processing the prediction residual error of the charging model by adopting a sliding window analysis method, and determining a good fault early warning threshold value. And finally, applying the charging prediction model meeting the requirements and the determined early warning threshold value to the real-time monitoring of the high-power direct-current charging of the electric automobile, and realizing the fault early warning of the high-power direct-current charging process of the electric automobile.
In order to achieve the above purpose, the invention provides the following scheme: the method for monitoring the high-power direct-current charging state of the electric automobile and early warning faults based on deep learning specifically comprises the following steps:
step 1: monitoring various parameters of the high-power direct-current charging process of the electric automobile in a state, and storing monitoring data into a database;
step 2: dividing the direct current charging data in the database into historical data and real-time data, and preprocessing the historical data and the real-time data;
and step 3: designing a CNN-BiGRU deep learning model to fully learn normal direct-current charging historical data and constructing a high-power direct-current charging prediction model of the electric automobile;
and 4, step 4: optimizing the hyper-parameters of the high-power direct-current charging prediction model of the electric automobile by adopting a sparrow search algorithm;
and 5: an evaluation standard of the output precision of the prediction model is formulated for evaluating the accuracy of the prediction model;
step 6: residual error analysis is carried out on the model predicted value through a sliding window method, and a fault early warning threshold value and a fault early warning rule which are suitable for direct-current charging of the electric automobile are determined;
and 7: acquiring real-time high-power direct-current charging data of the electric automobile on line;
and 8: inputting real-time high-power direct-current charging data into a trained prediction model to obtain a prediction output value;
and step 9: calculating residual mean and standard deviation of the predicted output value by a sliding window method;
step 10: and when the residual mean value and the standard deviation exceed the set threshold value at the same time, carrying out fault early warning and stopping the high-power direct current charging of the electric automobile.
In step 1, the state monitoring of each parameter in the high-power direct-current charging process of the electric vehicle includes, but is not limited to, parameter information such as rated capacity of a power battery of the whole vehicle, rated voltage of the power battery of the whole vehicle, maximum allowable monomer voltage, maximum allowable direct-current charging current, nominal total energy of the power battery of the whole vehicle, maximum allowable direct-current charging voltage, maximum allowable temperature, initial SOC of the power battery of the whole vehicle, initial voltage of the power battery of the whole vehicle, direct-current voltage required by the power battery of the whole vehicle, direct-current required by the power battery of the whole vehicle, measured value of the direct-current charging voltage, maximum monomer voltage of the power battery of the whole vehicle, current SOC of the power battery of the whole vehicle, and maximum temperature of the power battery monomer of the whole vehicle.
In step 2 of the invention, the high-power direct current charging data is preprocessed, which specifically comprises the following operations:
(1) performing outlier detection on the data, and deleting abnormal data in the data;
(2) filling missing values in the data by an interpolation method;
(3) and (4) normalizing the data by using a range normalization method, wherein the range of the processed data is [0,1 ].
The invention designs a CNN-BiGRU deep learning model in step 3, wherein the CNN has the following calculation formula:
ct＝f(WCNN*nt+bCNN)
in the formula, WCNNRepresenting a weight coefficient of a filter in the convolution of the high-power direct-current charging data of the electric automobile, namely a convolution kernel; n istRepresenting the high-power direct current charging data of the electric automobile at the time t; is a convolution operation; bCNNThe deviation coefficient represents the convolution operation of the high-power direct-current charging data of the electric automobile; c. CtThe method comprises the steps of obtaining a high-power direct-current charging data sequence of the electric vehicle, which is extracted after convolution; f represents an activation function of convolution operation of the high-power direct-current charging data of the electric automobile.
In step 3 of the invention, a CNN-BiGRU deep learning model is designed, wherein the BiGRU consists of a forward hidden gate control circulation Unit (GRU) and a backward hidden gate control circulation Unit (GRU), and the calculation formula of the GRU is as follows:
rt＝σ(Wrxt+Urht-1+br)
zt＝σ(Wzxt+Uzht-1+bz)
ht＝tanh(Wh1xt+(rt⊙ht-1)Wh2+bh)
ht＝(1-zt)⊙ht-1+zt⊙ht
yt＝σ(Wo⊙ht)
in the formula, rtTo reset the gate; z is a radical oftTo update the door; h istThe GRU hidden state is used for charging the electric automobile with high-power direct current at the time t; y istPredicting and outputting the high-power direct current charging of the electric automobile at the time t; σ and tanh are activation functions; wr、Wz、Ur、Uz、Wh1And Wh2Weighting parameters of deep features of the high-power direct-current charging data of the electric vehicle; br、bzAnd bhDeviation parameters of deep characteristics of the high-power direct-current charging data of the electric automobile are obtained; as a Hadamard product; h istAnd the GRU candidate state is a GRU candidate state for the high-power direct-current charging of the electric automobile at the time t. The calculation formula of BiGRU is as follows:
in the formula, wtThe output weight of a forward hidden layer GRU of deep characteristics of the high-power direct-current charging data of the electric vehicle at the time t; v. oftOutputting weight of backward hidden layer GRU of deep layer characteristics of the high-power direct current charging data of the electric vehicle at the time t; h istThe hidden state of the BiGRU is charged for the high-power direct current of the electric vehicle at the moment t; btIs htThe corresponding offset.
In step 4, the hyper-parameters of the model are optimized by adopting a sparrow search algorithm, the sparrow population in the sparrow search algorithm is divided into discoverers and followers, a certain proportion of sparrows are selected from the population for detection and early warning, and food is abandoned when danger occurs. When the number of sparrows is N and the searched space is D, the position updating formula of the discoverer is as follows:
in the formula, i and j represent the position information of the ith sparrow in the jth dimension; itermaxExpressed as the maximum number of iterations; q is represented as a normally distributed random number; l is represented as a 1 × d matrix and its elements are all 1. The follower's update formula is:
in the formula, xworstIs the current global worst position; x is the number ofpThe optimal position for the finder to occupy is a 1 x d matrix, each element in the matrix is randomly assigned a value of 1 or-1, and A+＝AT(AAT)-1. The sparrows with the detection and early warning mechanism behaviors account for 10% -20% of the total quantity, the initial positions of the sparrows are randomly generated in the population, and the position formula is as follows:
in the formula (I), the compound is shown in the specification,
the current global optimal position is obtained; beta is a normal distribution random number with the average value of 0 and the variance of 1, and is a step length control parameter; k represents the direction of movement of the sparrows and is also a step-size control parameter, the value of which is [ -1,1]A random number in between; f. ofiFitness of current sparrow individualsA value; f. ofgIs the current global optimum fitness value; f. ofwIs the current global worst fitness value; ε is a constant that prevents the denominator from appearing zero.
In the step 5 of the invention, the evaluation standard of the model prediction precision is formulated by adopting the root mean square error eRMSE(Root Mean Square Error, RMSE) and the Mean absolute percent Error eMAPETwo Error measurement modes (Mean Absolute percent Error, MAPE) are used as indexes for evaluating the accuracy of the direct current charging prediction model, and the calculation formula is as follows:
in the formula, yiAnd
respectively obtaining an actual value and a predicted value of the high-power direct current charging of the electric automobile at the ith moment; n is the number of all samples as a test set. e.g. of the typeRMSEAnd eMAPEThe smaller the value is, the more accurate the predicted high-power direct-current charging data of the electric automobile is.
In step 6 of the invention, residual error analysis is carried out on the model predicted value through the sliding window, and a proper fault early warning threshold value and rule are determined, so that the influence of error direct current charging data on residual error change in the data transmission process can be eliminated, and the error early warning can be effectively avoided. When the width of the sliding window is N, the mean value of the residual errors under the window
And the standard deviation S is calculated as follows:
in the formula, eiIs the residual error of the ith sample point in the sliding window. Analyzing and processing the residual error of the normal direct current charging data by utilizing a sliding window to obtain the maximum value of the average absolute value of the normal direct current charging residual errors
And the maximum value S of the residual standard deviationmaxThe calculation formula of the early warning threshold value is
SY＝k2Smax
In the formula, k1And k2The value of the scaling factor is determined by the model of the high-power direct-current charged electric automobile and the battery capacity. And when the mean value and the standard deviation both exceed the calculated early warning threshold value, carrying out fault early warning.
In step 10, the fault early warning in the high-power direct-current charging process of the electric automobile is realized, and when the residual mean value and the standard deviation exceed the set threshold value at the same time, the fault early warning is carried out, and the high-power direct-current charging of the electric automobile is cut off, so that the fire accident is prevented.
The beneficial effect of this application lies in: according to the method for monitoring the high-power direct-current charging state and early warning the fault of the electric automobile based on deep learning, after the high-power direct-current charging data of the electric automobile are preprocessed, a CNN-BiGRU deep learning model is designed to deeply learn the high-power direct-current charging data of the electric automobile, and a normal high-power direct-current charging prediction model of the electric automobile is constructed; according to the method and the device, the CNN is adopted to carry out deep excavation on the high-power direct-current charging data, the deep features of the direct-current charging data are extracted, and the time sequence analysis is carried out on the deep features by utilizing the advantages of the BiGRU analysis history and the future data, so that the training time of the model is shortened, and the prediction accuracy of the model is improved. The method and the device use the sparrow search algorithm to determine the hyper-parameters of the CNN-BiGRU deep learning model, and can further enhance the accuracy of predicting the high-power direct-current charging data of the electric vehicle; the method and the device for determining the fault early warning of the electric automobile adopt a sliding window analysis method to determine the rule and the threshold value of the fault early warning of the electric automobile, not only can early warn the fault of the high-power direct-current charging process of the electric automobile in advance, but also can eliminate the error early warning caused by error data in the data transmission process.
Drawings
FIG. 1 is a schematic flow chart of a deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to the invention;
FIG. 2 is a structural diagram of a CNN-BiGRU deep learning model designed by the present invention;
FIG. 3 is a flow chart of parameter optimization of a CNN-BiGRU deep learning model based on a sparrow search algorithm according to the present invention;
FIG. 4 is a block diagram of a deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method.
Detailed Description
Specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings of the specification. The procedures, conditions, experimental methods and the like for carrying out the present invention are general knowledge and common general knowledge in the art except for the contents specifically mentioned below, and the present invention is not particularly limited.
FIG. 1 is a schematic flow chart of a deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method. As shown in fig. 1, the method for monitoring the high-power dc charging state and early warning the fault of the electric vehicle based on deep learning of the present invention comprises the following steps:
step 1: the method comprises the steps of carrying out state monitoring on various parameters in the high-power direct-current charging process of the electric automobile, and monitoring state information such as rated capacity of a power battery of the whole automobile, rated voltage of the power battery of the whole automobile, maximum allowable single voltage, maximum allowable direct-current charging current, rated total energy of the power battery of the whole automobile, maximum allowable direct-current charging voltage, maximum allowable temperature, initial SOC of the power battery of the whole automobile, initial voltage of the power battery of the whole automobile, direct-current voltage required by the power battery of the whole automobile, direct-current charging voltage measured value, direct-current charging current measured value, maximum single voltage of the power battery of the whole automobile, current SOC of the power battery of the whole automobile, highest single temperature of the power battery of the whole automobile and the like.
Step 2: and dividing the data set into historical direct current charging data and real-time direct current charging data, and preprocessing the historical direct current charging data and the real-time direct current charging data. Specifically, historical normal direct current charging data is used for constructing a high-power direct current normal charging prediction model of the electric automobile, and real-time direct current charging data is used for online fault early warning.
The preprocessing of the data comprises the following operations:
(1) performing outlier detection on the data, and deleting abnormal data in the data;
(2) filling missing values in the data by an interpolation method;
(3) and (3) carrying out normalization processing on the data by using a range normalization method, wherein the processed data range is [0,1], and a calculation formula is as follows.
In the formula, xmin，xmaxRespectively, the minimum and maximum values, x, of the data set sample homogeneous dataoutIs the result of normalizing the input data x.
And step 3: and designing a CNN-BiGRU deep learning model, fully learning the normal direct current charging historical data of the electric automobile, and constructing a high-power direct current charging prediction model of the electric automobile.
The CNN network structure in the CNN-BiGRU deep learning model is shown in the CNN network structure in fig. 2, and the calculation formula is as follows:
ct＝f(WCNN*nt+bCNN)
in the formula, WCNNRepresenting the weight coefficient of the filter in the convolution of the high-power DC charging data of the electric vehicle, i.e. the volumeAccumulating kernels; n istRepresenting the high-power direct current charging data of the electric automobile at the time t; is a convolution operation; bCNNThe deviation coefficient represents the convolution operation of the high-power direct-current charging data of the electric automobile; c. CtThe method comprises the steps of obtaining a high-power direct-current charging data sequence of the electric vehicle, which is extracted after convolution; f represents an activation function of convolution operation of the high-power direct-current charging data of the electric automobile.
The BiGRU is transformed based on the GRU, has strong memory capacity, can effectively retain historical input data, and compared with the unidirectional GRU, the BiGRU can give consideration to the influence of historical and future charging data on the current moment, so that deep analysis can be performed on the historical high-power direct-current charging data of the electric automobile. The CNN-BiGRU deep learning model has the advantages of both CNN and BiGRU networks, and the model structure is shown in FIG. 2.
GRU update gate ztAnd a reset gate rtAnd the updating door represents the influence degree of the charging data of the electric automobile at the previous moment on the current moment. The reset gate represents the degree of neglected charging data of the electric vehicle at the previous moment, and the specific calculation formula is as follows:
rt＝σ(Wrxt+Urht-1+br)
zt＝σ(Wzxt+Uzht-1+bz)
ht＝tanh(Wh1xt+(rt⊙ht-1)Wh2+bh)
ht＝(1-zt)⊙ht-1+zt⊙ht
yt＝σ(Wo⊙ht)
in the formula, rtTo reset the gate; z is a radical oftTo update the door; h istThe GRU hidden state is used for charging the electric automobile with high-power direct current at the time t; y istPredicting and outputting the high-power direct current charging of the electric automobile at the time t; σ and tanh are activation functions; wr、Wz、Ur、Uz、Wh1And Wh2Is electricityWeighting parameters of deep features of the high-power direct-current charging data of the electric vehicle; br、bzAnd bhDeviation parameters of deep characteristics of the high-power direct-current charging data of the electric automobile are obtained; as a Hadamard product; h istGRU candidate state for charging electric vehicle with high-power direct current at time t and formed by reset gate rtGRU hidden state h for high-power direct-current charging of electric automobile at time t-1t-1And the input x of the high-power direct current charging of the electric automobile at the current momenttAnd (4) controlling together.
As shown in the BiGRU network structure in FIG. 2, the BiGRU is composed of a forward hidden layer GRU and a backward hidden layer GRU which are connected with each other, and the charging output of the electric vehicle at the time t and the output of the forward hidden layer GRU
And the output of backward hidden layer GRU
The result of the linear superposition is calculated as follows:
in the formula, wtThe output weight of the GRU of the forward hidden layer of the deep characteristics of the high-power direct-current charging data of the electric vehicle at the moment t; v. oftOutputting the weight of the backward hidden layer GRU of the deep characteristics of the high-power direct-current charging data of the electric vehicle at the time t; h istThe hidden state of the BiGRU is charged for the high-power direct current of the electric vehicle at the moment t; btIs htThe corresponding offset.
And 4, step 4: and optimizing the hyper-parameters of the CNN-BiGRU deep learning model by adopting a sparrow search algorithm. The CNN-BiGRU deep learning model sets different learning rates, iteration times and the number of hidden layer units, and the obtained prediction performance has larger difference. The structure of the current CNN-BiGRU deep learning model is usually obtained by selecting a plurality of groups of different learning rates, iteration times and the number of hidden layer units for debugging and comparison according to experience, which usually needs to consume a large amount of time and energy. Aiming at the problem, the invention provides a method for solving hyper-parameters of a model by using a sparrow search algorithm, and a flow chart is shown in FIG. 3 and specifically comprises the following steps:
step 4.1: initializing a sparrow population, and determining the optimizing dimension of sparrows according to the hyper-parameters;
step 4.2: and (3) evaluating the positions of the initialized sparrows according to a fitness function and sequencing, wherein the first 20% of the sparrows are discoverers, the rest are followers, and 10% -20% of the sparrows are randomly selected to carry a reconnaissance early warning mechanism to act. The location update formula of the discoverer is as follows:
in the formula, i and j represent the position information of the ith sparrow in the jth dimension; itermaxExpressed as the maximum number of iterations; q is represented as a normally distributed random number; l is represented as a 1 × d matrix and its elements are all 1. The follower's update formula is:
in the formula, xworstIs the current global worst position; x is the number ofpThe optimal position for the finder to occupy is a 1 x d matrix, each element in the matrix is randomly assigned a value of 1 or-1, and A+＝AT(AAT)-1. The position formula of the sparrows with the detection and early warning mechanism behaviors is as follows:
in the formula (I), the compound is shown in the specification,
the current global optimal position is obtained; beta is a normal distribution random number with the average value of 0 and the variance of 1, and is a step length control parameter; k represents the direction of movement of the sparrows and is also a step-size control parameter, the value of which is [ -1,1]A random number in between; f. ofiThe fitness value of the current sparrow individual is obtained; f. ofgIs the current global optimum fitness value; f. ofwIs the current global worst fitness value; ε is a constant that prevents the denominator from appearing zero.
Step 4.3: and updating the sparrow position by using the discoverer, the follower and the investigation early warning formula, constraining the hyper-parameters by using a boundary function, transmitting the hyper-parameters to the CNN-BiGRU deep learning model for prediction, and returning the result to evaluate the position by using a fitness function.
Step 4.4: if the fitness of the current sparrow position is better than that of the best position, replacing, otherwise, keeping unchanged;
step 4.5: if the optimal fitness of the sparrows in the iteration is better than the global optimal fitness, replacing, otherwise, keeping unchanged;
step 4.6: and if the error does not reach the set value, returning to the step 4.3, updating the positions of all the sparrow populations, and recalculating until the maximum iteration number is reached. The whole process is shown in fig. 3.
And 5: making an evaluation standard of model prediction accuracy by adopting eRMSEAnd eMAPEThe two error measurement modes are used as indexes for evaluating the accuracy of the direct current charging prediction model, and the calculation formula is as follows:
in the formula, yiAnd
respectively obtaining an actual value and a predicted value of the high-power direct current charging of the electric automobile at the ith moment; n is the number of all samples as a test set. e.g. of the typeRMSEAnd eMAPEThe smaller the value is, the more accurate the predicted high-power direct-current charging data of the electric automobile is.
Step 6: residual error analysis is carried out on the model predicted value through the sliding window, a proper fault early warning threshold value and a proper fault early warning rule are determined, the influence of error direct current charging data on residual error change in the data transmission process can be eliminated, and error early warning can be effectively avoided. When the width of the sliding window is N, the calculation formula of the mean value and the standard deviation of the residual error under the window is as follows:
in the formula, eiIs the residual error of the ith sample point in the sliding window. Analyzing and processing the residual error of the normal direct current charging data by utilizing a sliding window to obtain the maximum value of the average absolute value of the normal direct current charging residual errors
And the maximum value S of the residual standard deviationmaxThe calculation formula of the early warning threshold value is
SY＝k2Smax
In the formula, k1And k2The proportional coefficient is the electric automobile model and the battery capacity which are charged by high-power direct currentAnd (4) determining. And when the mean value and the standard deviation both exceed the calculated early warning threshold value, carrying out fault early warning.
And 7: acquiring real-time high-power direct-current charging data of the electric automobile on line;
and 8: inputting real-time high-power direct-current charging data into a trained prediction model to obtain a prediction output value;
and step 9: calculating residual mean and standard deviation of the predicted output value by a sliding window method;
step 10: and when the residual mean value and the standard deviation exceed the set threshold value at the same time, carrying out fault early warning and stopping the high-power direct current charging of the electric automobile.
Although the present invention has been disclosed in the preferred embodiments above, the present invention is not limited thereto. Various changes and modifications may be effected therein by one of ordinary skill in the pertinent art without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention is subject to the scope defined by the claims.
Claims (10)
1. The method for monitoring the high-power direct-current charging state and early warning the fault of the electric automobile based on deep learning is characterized by comprising three parts of state monitoring, data preprocessing, offline model training and online fault early warning, and specifically comprises the following steps:
step 1: monitoring various parameters of the high-power direct-current charging process of the electric automobile in a state, and storing monitoring data into a database;
step 2: dividing the direct current charging data in the database into historical data and real-time data, and preprocessing the historical data and the real-time data;
and step 3: designing a CNN-BiGRU deep learning model, fully learning the normal direct current charging historical data of the electric automobile, and constructing a high-power direct current charging prediction model of the electric automobile;
and 4, step 4: optimizing the hyper-parameters of the high-power direct-current charging prediction model of the electric automobile by adopting a sparrow search algorithm;
and 5: an evaluation standard of the output precision of the prediction model is formulated for evaluating the accuracy of the prediction model;
step 6: residual error analysis is carried out on the predicted value of the model through a sliding window method, and a fault early warning threshold value and a fault early warning rule which are suitable for direct-current charging of the electric automobile are determined;
and 7: acquiring real-time high-power direct-current charging data of the electric automobile on line;
and 8: inputting real-time high-power direct-current charging data into a trained prediction model to obtain a prediction output value;
and step 9: calculating residual mean and standard deviation of the predicted output value by a sliding window method;
step 10: and when the residual mean value and the standard deviation exceed the set threshold value at the same time, carrying out fault early warning and stopping the high-power direct current charging of the electric automobile.
2. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, the method is characterized in that various parameters for monitoring the state of the high-power direct-current charging of the electric automobile in the step 1 include, but are not limited to, rated capacity of a power battery of the whole automobile, rated voltage of the power battery of the whole automobile, maximum allowable single voltage, maximum allowable direct-current charging current, nominal total energy of the power battery of the whole automobile, maximum allowable direct-current charging voltage, maximum allowable temperature, initial SOC of the power battery of the whole automobile, initial voltage of the power battery of the whole automobile, direct-current voltage required by the power battery of the whole automobile, direct-current required by the power battery of the whole automobile, measured value of the direct-current charging voltage, maximum single voltage of the power battery of the whole automobile, current SOC of the power battery of the whole automobile, maximum temperature of the power battery of the whole automobile and the like.
3. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, wherein the step 2 is used for preprocessing direct-current charging data, and the method specifically comprises the following operations:
(1) performing outlier detection on the data, and deleting abnormal data in the data;
(2) filling missing values in the data by an interpolation method;
(3) and (4) normalizing the data by using a range normalization method, wherein the range of the processed data is [0,1 ].
4. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, wherein the CNN-BiGRU deep learning model designed in the step 3 is a Convolutional Neural Network (CNN), and a calculation formula of the CNN is as follows:
ct＝f(WCNN*nt+bCNN)
in the formula, WCNNRepresenting a weight coefficient of a filter in the convolution of the high-power direct-current charging data of the electric automobile, namely a convolution kernel; n istRepresenting the high-power direct current charging data of the electric automobile at the time t; is a convolution operation; bCNNThe deviation coefficient represents the convolution operation of the high-power direct-current charging data of the electric automobile; c. CtThe method comprises the steps of obtaining a high-power direct-current charging data sequence of the electric vehicle, which is extracted after convolution; f represents an activation function of convolution operation of the high-power direct-current charging data of the electric automobile.
5. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, wherein the CNN-BiGRU deep learning model designed in step 3 is a bidirectional gating cycle Unit, which is composed of a forward and a backward hidden layer gating cycle Unit (GRU), and the calculation formula of the GRU is as follows:
rt＝σ(Wrxt+Urht-1+br)
zt＝σ(Wzxt+Uzht-1+bz)
ht＝tanh(Wh1xt+(rt⊙ht-1)Wh2+bh)
ht＝(1-zt)⊙ht-1+zt⊙ht
yt＝σ(Wo⊙ht)
in the formula, rtTo reset the gate; z is a radical oftTo update the door; h istThe GRU hidden state is used for charging the electric automobile with high-power direct current at the time t; y istPredicting and outputting the high-power direct current charging of the electric automobile at the time t; σ and tanh are activation functions; wr、Wz、Ur、Uz、Wh1And Wh2Weighting parameters of deep features of the high-power direct-current charging data of the electric vehicle; br、bzAnd bhDeviation parameters of deep characteristics of the high-power direct-current charging data of the electric automobile are obtained; as a Hadamard product; h istAnd the GRU candidate state is a GRU candidate state for the high-power direct-current charging of the electric automobile at the time t. The calculation formula of BiGRU is as follows:
in the formula, wtThe output weight of a forward hidden layer GRU of deep characteristics of the high-power direct-current charging data of the electric vehicle at the time t; v. oftOutputting weight of backward hidden layer GRU of deep layer characteristics of the high-power direct current charging data of the electric vehicle at the time t; h istThe hidden state of the BiGRU is charged for the high-power direct current of the electric vehicle at the moment t; btIs htThe corresponding offset.
6. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, characterized in that a sparrow search algorithm is adopted in the step 4 to optimize hyper-parameters of a prediction model, sparrow populations in the sparrow search algorithm are divided into discoverers and followers, a certain proportion of sparrows in the populations are selected for detection and early warning, and food is given up when danger occurs. When the number of sparrows is N and the searched space is D, the position updating formula of the discoverer is as follows:
in the formula, i and j represent the position information of the ith sparrow in the jth dimension; itermaxExpressed as the maximum number of iterations; q is represented as a normally distributed random number; l is represented as a 1 × d matrix and its elements are all 1. The follower's update formula is:
in the formula, xworstIs the current global worst position; x is the number ofpThe optimal position for the finder to occupy is a 1 x d matrix, each element in the matrix is randomly assigned a value of 1 or-1, and A+＝AT(AAT)-1. The sparrows with the detection and early warning mechanism behaviors account for 10% -20% of the total quantity, the initial positions of the sparrows are randomly generated in the population, and the position formula is as follows:
in the formula (I), the compound is shown in the specification,
the current global optimal position is obtained; beta is a step length control parameter, and is a normally distributed random number with the average value of 0 and the variance of 1; k represents the direction of movement of the sparrows and is also a step-size control parameter, the value of which is [ -1,1]A random number in between; f. ofiAdaptation to current sparrow individualsA value of the metric; f. ofgIs the current global optimum fitness value; f. ofwIs the current global worst fitness value; ε is a constant that prevents the denominator from appearing zero.
7. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method as claimed in claim 1, wherein the evaluation standard of the output accuracy of the prediction model formulated in the step 5 is that the root mean square error e is adoptedRMSE(Root Mean Square Error, RMSE) and the Mean absolute percent Error eMAPETwo Error measurement modes (Mean Absolute percent Error, MAPE) are used as indexes for evaluating the accuracy of the direct current charging prediction model, and the calculation formula is as follows:
in the formula, yiAnd
respectively obtaining an actual value and a predicted value of the high-power direct current charging of the electric automobile at the ith moment; n is the number of all samples as a test set. e.g. of the typeRMSEAnd eMAPEThe smaller the value is, the more accurate the predicted high-power direct-current charging data of the electric automobile is.
8. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, wherein the model predicted value is subjected to residual error analysis through a sliding window method in the step 6, and a proper fault early warning threshold value and rule are determined, so that the influence of wrong direct-current charging data on residual error change in a data transmission process can be eliminated, and the false early warning can be effectively avoided. When the width of the sliding window is N, the windowMean of lower residuals
And the standard deviation S is calculated as follows:
in the formula, eiIs the residual error of the ith sample point in the sliding window. Analyzing and processing the residual error of the normal direct current charging data by utilizing a sliding window to obtain the maximum value of the average absolute value of the normal direct current charging residual errors
And the maximum value S of the residual standard deviationmaxThe calculation formula of the early warning threshold value is
SY＝k2Smax
In the formula, k1And k2The value of the scaling factor is determined by the model of the high-power direct-current charged electric automobile and the battery capacity. And when the mean value and the standard deviation both exceed the calculated early warning threshold value, carrying out fault early warning.
9. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method according to claim 1, wherein fault early warning is realized in the step 10 in the electric vehicle high-power direct-current charging process, and when the residual mean value and the standard deviation simultaneously exceed the set threshold, fault early warning is performed, and the high-power direct-current charging of the electric vehicle is cut off, so that fire accidents are prevented.
10. The deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method as claimed in claim 1, wherein the step 10 of high-power direct-current charging of the electric vehicle means that the charging power is greater than 120kW and less than 360 kW.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111203154.7A CN113895271A (en) | 2021-10-15 | 2021-10-15 | Deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111203154.7A CN113895271A (en) | 2021-10-15 | 2021-10-15 | Deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113895271A true CN113895271A (en) | 2022-01-07 |
Family
ID=79192252
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111203154.7A Withdrawn CN113895271A (en) | 2021-10-15 | 2021-10-15 | Deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113895271A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115447439A (en) * | 2022-09-19 | 2022-12-09 | 特瓦特能源科技有限公司 | Charging safety early warning method, system, equipment and medium based on battery temperature |
| CN115891741A (en) * | 2022-09-30 | 2023-04-04 | 南京邮电大学 | Remote fault early warning method and device suitable for electric vehicle charging process |
| CN116552299A (en) * | 2023-07-11 | 2023-08-08 | 深圳市南霸科技有限公司 | Movable electric automobile emergency charging system and method |
-
2021
- 2021-10-15 CN CN202111203154.7A patent/CN113895271A/en not_active Withdrawn
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115447439A (en) * | 2022-09-19 | 2022-12-09 | 特瓦特能源科技有限公司 | Charging safety early warning method, system, equipment and medium based on battery temperature |
| CN115891741A (en) * | 2022-09-30 | 2023-04-04 | 南京邮电大学 | Remote fault early warning method and device suitable for electric vehicle charging process |
| CN115891741B (en) * | 2022-09-30 | 2023-09-22 | 南京邮电大学 | Remote fault early warning method and device suitable for electric automobile charging process |
| CN116552299A (en) * | 2023-07-11 | 2023-08-08 | 深圳市南霸科技有限公司 | Movable electric automobile emergency charging system and method |
| CN116552299B (en) * | 2023-07-11 | 2023-09-15 | 深圳市南霸科技有限公司 | Movable electric automobile emergency charging system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113895271A (en) | Deep learning-based electric vehicle high-power direct-current charging state monitoring and fault early warning method | |
| CN112986827B (en) | Fuel cell residual life prediction method based on deep learning | |
| CN113895272A (en) | Electric vehicle alternating current charging state monitoring and fault early warning method based on deep learning | |
| CN111813084B (en) | Mechanical equipment fault diagnosis method based on deep learning | |
| CN110689171A (en) | Turbine health state prediction method based on E-LSTM | |
| CN111680848A (en) | Battery life prediction method based on prediction model fusion and storage medium | |
| CN113705615B (en) | Multi-stage equipment fault diagnosis method and system for electric automobile charging process based on neural network | |
| CN114084024A (en) | Electric automobile high-power charging monitoring and multi-stage pre-alarming method based on charging network | |
| CN116541790B (en) | New energy vehicle health assessment method and device based on multi-feature fusion | |
| CN115366683A (en) | Fault diagnosis strategy for new energy automobile power battery multi-dimensional model fusion | |
| CN114925612A (en) | Transformer fault diagnosis method for optimizing hybrid kernel extreme learning machine based on sparrow search algorithm | |
| CN111983474A (en) | Lithium ion battery life prediction method and system based on capacity decline model | |
| CN112882909A (en) | Fuel cell system fault prediction method and device | |
| CN113807594A (en) | IVMD-SSA-LSSVM-based power load prediction method | |
| CN115236522A (en) | End-to-end capacity estimation method of energy storage battery based on hybrid deep neural network | |
| CN116125306A (en) | Power battery thermal runaway diagnosis method and system based on hybrid prediction framework | |
| CN112734131A (en) | Fan blade icing state prediction method based on deep learning algorithm | |
| CN114280490B (en) | Lithium ion battery state of charge estimation method and system | |
| CN114036647A (en) | Power battery safety risk assessment method based on real vehicle data | |
| CN114460481A (en) | Energy storage battery thermal runaway early warning method based on Bi-LSTM and attention mechanism | |
| CN116653608B (en) | Electric automobile charging protection and control method, device and storage medium | |
| CN116680598A (en) | Fault diagnosis and residual life prediction method for intelligent distribution box of construction site | |
| CN117113202A (en) | Power loop energy consumption detection method and equipment based on joint error stacking model | |
| CN115865716A (en) | Network state analysis method, system and computer readable medium | |
| CN115828744A (en) | White light LED fault on-line diagnosis and service life prediction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20220107 |